/*********************************************************************/
/* Copyright 2009, 2010 The University of Texas at Austin. */
/* All rights reserved. */
/* */
/* Redistribution and use in source and binary forms, with or */
/* without modification, are permitted provided that the following */
/* conditions are met: */
/* */
/* 1. Redistributions of source code must retain the above */
/* copyright notice, this list of conditions and the following */
/* disclaimer. */
/* */
/* 2. Redistributions in binary form must reproduce the above */
/* copyright notice, this list of conditions and the following */
/* disclaimer in the documentation and/or other materials */
/* provided with the distribution. */
/* */
/* THIS SOFTWARE IS PROVIDED BY THE UNIVERSITY OF TEXAS AT */
/* AUSTIN ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, */
/* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
/* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
/* DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT */
/* AUSTIN OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, */
/* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES */
/* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE */
/* GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR */
/* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF */
/* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT */
/* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT */
/* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE */
/* POSSIBILITY OF SUCH DAMAGE. */
/* */
/* The views and conclusions contained in the software and */
/* documentation are those of the authors and should not be */
/* interpreted as representing official policies, either expressed */
/* or implied, of The University of Texas at Austin. */
/*********************************************************************/
#define ASSEMBLER
#include "common.h"
#define M ARG1 /* rdi */
#define X ARG2 /* rsi */
#define INCX ARG3 /* rdx */
#define I %rax
#include "l1param.h"
PROLOGUE
PROFCODE
SAVEREGISTERS
xorps %xmm0, %xmm0
testq M, M
jle .L999
testq INCX, INCX
jle .L999
xorps %xmm1, %xmm1
xorps %xmm2, %xmm2
xorps %xmm3, %xmm3
pcmpeqb %xmm15, %xmm15
psrlq $1, %xmm15
salq $ZBASE_SHIFT, INCX
cmpq $2 * SIZE, INCX
jne .L40
subq $-16 * SIZE, X
addq M, M
testq $SIZE, X
je .L05
#ifdef movsd
xorps %xmm0, %xmm0
#endif
movsd -16 * SIZE(X), %xmm0
addq $SIZE, X
andps %xmm15, %xmm0
subq $1, M
jle .L999
ALIGN_3
.L05:
movq M, I
sarq $4, I
jle .L20
movaps -16 * SIZE(X), %xmm4
movaps -14 * SIZE(X), %xmm5
movaps -12 * SIZE(X), %xmm6
movaps -10 * SIZE(X), %xmm7
movaps -8 * SIZE(X), %xmm8
movaps -6 * SIZE(X), %xmm9
movaps -4 * SIZE(X), %xmm10
movaps -2 * SIZE(X), %xmm11
decq I
jle .L11
ALIGN_4
.L10:
#ifdef PREFETCH
PREFETCH (PREFETCHSIZE + 0) - PREOFFSET(X)
#endif
andps %xmm15, %xmm4
addpd %xmm4, %xmm0
movaps 0 * SIZE(X), %xmm4
andps %xmm15, %xmm5
addpd %xmm5, %xmm1
movaps 2 * SIZE(X), %xmm5
andps %xmm15, %xmm6
addpd %xmm6, %xmm2
movaps 4 * SIZE(X), %xmm6
andps %xmm15, %xmm7
addpd %xmm7, %xmm3
movaps 6 * SIZE(X), %xmm7
#if defined(PREFETCH) && !defined(FETCH128)
PREFETCH (PREFETCHSIZE + 64) - PREOFFSET(X)
#endif
andps %xmm15, %xmm8
addpd %xmm8, %xmm0
movaps 8 * SIZE(X), %xmm8
andps %xmm15, %xmm9
addpd %xmm9, %xmm1
movaps 10 * SIZE(X), %xmm9
andps %xmm15, %xmm10
addpd %xmm10, %xmm2
movaps 12 * SIZE(X), %xmm10
andps %xmm15, %xmm11
addpd %xmm11, %xmm3
movaps 14 * SIZE(X), %xmm11
subq $-16 * SIZE, X
decq I
jg .L10
ALIGN_4
.L11:
andps %xmm15, %xmm4
andps %xmm15, %xmm5
andps %xmm15, %xmm6
andps %xmm15, %xmm7
addpd %xmm4, %xmm0
addpd %xmm5, %xmm1
addpd %xmm6, %xmm2
addpd %xmm7, %xmm3
andps %xmm15, %xmm8
andps %xmm15, %xmm9
andps %xmm15, %xmm10
andps %xmm15, %xmm11
addpd %xmm8, %xmm0
addpd %xmm9, %xmm1
addpd %xmm10, %xmm2
addpd %xmm11, %xmm3
subq $-16 * SIZE, X
ALIGN_3
.L20:
andq $15, M
jle .L998
testq $8, M
je .L21
movaps -16 * SIZE(X), %xmm4
movaps -14 * SIZE(X), %xmm5
movaps -12 * SIZE(X), %xmm6
movaps -10 * SIZE(X), %xmm7
andps %xmm15, %xmm4
andps %xmm15, %xmm5
andps %xmm15, %xmm6
andps %xmm15, %xmm7
addpd %xmm4, %xmm0
addpd %xmm5, %xmm1
addpd %xmm6, %xmm2
addpd %xmm7, %xmm3
addq $8 * SIZE, X
ALIGN_3
.L21:
testq $4, M
je .L22
movaps -16 * SIZE(X), %xmm4
movaps -14 * SIZE(X), %xmm5
andps %xmm15, %xmm4
andps %xmm15, %xmm5
addpd %xmm4, %xmm0
addpd %xmm5, %xmm1
addq $4 * SIZE, X
ALIGN_3
.L22:
testq $2, M
je .L23
movaps -16 * SIZE(X), %xmm6
andps %xmm15, %xmm6
addpd %xmm6, %xmm3
addq $2 * SIZE, X
.L23:
testq $1, M
je .L998
#ifdef movsd
xorps %xmm4, %xmm4
#endif
movsd -16 * SIZE(X), %xmm4
andps %xmm15, %xmm4
addsd %xmm4, %xmm0
jmp .L998
ALIGN_3
.L40:
movq M, I
sarq $2, I
jle .L60
ALIGN_4
.L50:
#if defined(OPTERON) || defined(BARCELONA) || defined(SHANGHAI)
prefetcht0 PREFETCHSIZE * SIZE(X)
#endif
#ifdef PENTIUM4
prefetchnta PREFETCHSIZE * SIZE(X)
#endif
movsd 0 * SIZE(X), %xmm4
movhpd 1 * SIZE(X), %xmm4
addq INCX, X
andpd %xmm15, %xmm4
addpd %xmm4, %xmm0
movsd 0 * SIZE(X), %xmm5
movhpd 1 * SIZE(X), %xmm5
addq INCX, X
andpd %xmm15, %xmm5
addpd %xmm5, %xmm1
movsd 0 * SIZE(X), %xmm6
movhpd 1 * SIZE(X), %xmm6
addq INCX, X
andpd %xmm15, %xmm6
addpd %xmm6, %xmm2
movsd 0 * SIZE(X), %xmm7
movhpd 1 * SIZE(X), %xmm7
addq INCX, X
andpd %xmm15, %xmm7
addpd %xmm7, %xmm3
decq I
jg .L50
ALIGN_4
.L60:
andq $3, M
jle .L998
ALIGN_4
.L61:
movsd 0 * SIZE(X), %xmm4
movhpd 1 * SIZE(X), %xmm4
andpd %xmm15, %xmm4
addpd %xmm4, %xmm0
addq INCX, X
decq M
jg .L61
ALIGN_4
.L998:
addpd %xmm1, %xmm0
addpd %xmm3, %xmm2
addpd %xmm2, %xmm0
#ifndef HAVE_SSE3
movhlps %xmm0, %xmm1
addsd %xmm1, %xmm0
#else
haddpd %xmm0, %xmm0
#endif
ALIGN_4
.L999:
RESTOREREGISTERS
ret
EPILOGUE